这些表的内容在HIVE里面是有很格式化的显示,在HIVE里面输入describe formatted table_name;展示出来的信息就是表格中所有表(除了NUCLEUS_TABLE, SEQUENCE_TABLE这两张表)的信息,而且key值基本上和元数据的表字段名称一致。
HIVE如何定义和操作元数据:
定义:
\src\metastore\src\model\ package.jdo和org.apache.hadoop.hive.metastore.model.*用来定义元数据的数据字典和操作。
以TBLS表为例:
Mysql中的表结构为:
- mysql> desc TBLS;
- +--------------------+--------------+------+-----+---------+-------+
- | Field | Type | Null | Key | Default | Extra |
- +--------------------+--------------+------+-----+---------+-------+
- | TBL_ID | bigint(20) | NO | PRI | NULL | |
- | CREATE_TIME | int(11) | NO | | NULL | |
- | DB_ID | bigint(20) | YES | MUL | NULL | |
- | LAST_ACCESS_TIME | int(11) | NO | | NULL | |
- | OWNER | varchar(767) | YES | | NULL | |
- | RETENTION | int(11) | NO | | NULL | |
- | SD_ID | bigint(20) | YES | MUL | NULL | |
- | TBL_NAME | varchar(128) | YES | MUL | NULL | |
- | VIEW_EXPANDED_TEXT | mediumtext | YES | | NULL | |
- | VIEW_ORIGINAL_TEXT | mediumtext | YES | | NULL | |
- | TBL_TYPE | varchar(128) | YES | | NULL | |
- +--------------------+--------------+------+-----+---------+-------+
|
在package.jdo里的定义为:
- <<strong><span style="color: #0000ff;">class name="MTable"</span></strong> <strong><span style="color: #ff0000;">table="TBLS"</span></strong> identity-type="datastore" detachable="true">
- <datastore-identity>
- <column name="TBL_ID"/>
- </datastore-identity>
- <index name="UniqueTable" unique="true">
- <column name="TBL_NAME"/>
- <column name="DB_ID"/>
- </index>
- <field name="tableName">
- <column name="TBL_NAME" length="128" jdbc-type="VARCHAR"/>
- </field>
- <field name="database">
- <column name="DB_ID"/>
- </field>
- <field name="sd" dependent="true">
- <column name="SD_ID"/>
- </field>
- <field name="owner">
- <column name="OWNER" length="767" jdbc-type="VARCHAR"/>
- </field>
- <field name="createTime">
- <column name="CREATE_TIME" jdbc-type="integer"/>
- </field>
- <field name="lastAccessTime">
- <column name="LAST_ACCESS_TIME" jdbc-type="integer"/>
- </field>
- <field name="retention">
- <column name="RETENTION" jdbc-type="integer"/>
- </field>
- <field name="viewOriginalText">
- <column name="VIEW_ORIGINAL_TEXT" jdbc-type="LONGVARCHAR"/>
- </field>
- <field name="viewExpandedText">
- <column name="VIEW_EXPANDED_TEXT" jdbc-type="LONGVARCHAR"/>
- </field>
- <field name="tableType">
- <column name="TBL_TYPE" length="128" jdbc-type="VARCHAR"/>
- </field>
- </class>
|
表名为:TBLS;对应的类为:org.apache.hadoop.hive.metastore.model. MTable,TBLS的所有操作都在此类中定义。
操作:
org.apache.hadoop.hive.metastore. ObjectStore
定义了如何获取元数据,如何写入元数据,其实全部是调用javax.jdo中各个类的方法。
以listPartitionNames为例:
- public List<String> listPartitionNames(String dbName, String tableName,
- short max) throws MetaException {
- List<String> pns = new ArrayList<String>();
- boolean success = false;
- try {
- openTransaction();
- LOG.debug("Executing getPartitionNames");
- dbName = dbName.toLowerCase().trim();
- tableName = tableName.toLowerCase().trim();
- javax.jdo.Query q = pm.newQuery(
- "select partitionName from org.apache.hadoop.hive.metastore.model.MPartition "
- + "where table.database.name == t1 && table.tableName == t2 "
- + "order by partitionName asc");
- q.declareParameters("java.lang.String t1, java.lang.String t2");
- q.setResult("partitionName");
- Collection names = (Collection) q.execute(dbName, tableName);
- pns = new ArrayList<String>();
- for (Iterator i = names.iterator(); i.hasNext();) {
- pns.add((String) i.next());
- }
- success = commitTransaction();
- } finally {
- if (!success) {
- rollbackTransaction();
- }
- }
- return pns;
- }
|
所以,如果要修改HIVE元数据管理的功能,就可以按如下步骤操作:
1. 在\src\metastore\src\model\ package.jdo中定义新的HIVE元数据表的数据字典
2. 在org.apache.hadoop.hive.metastore.model中定义新增表的操作
3. 在org.apache.hadoop.hive.metastore. ObjectStore中新增方法处理如上定义的功能